** Description changed: - To read and scan SMBIOS in /dev/mem is dangerous and for DMI information - it will be better to read from /sys/firmware/dmi/tables/DMI if it - exists. + [Impact] - [278063.180820] Process lscpu (pid: 22668, stack limit = 0xffff842fb2a0c000) - [278063.187593] Stack: (0xffff842fb2a0fd90 to 0xffff842fb2a10000) - [278063.193412] fd80: ffff842fb2a0fe00 ffff00000829bf00 - [278063.201315] fda0: ffff801fae2a9a00 0000000025ccb150 0000000000000020 ffff842fb2a0feb0 - [278063.209217] fdc0: 0000000000000020 0000000000000015 0000000000000123 000000000000003f - [278063.217120] fde0: ffff000008a22000 ffff842faf449c00 ffff842fb2a0feb0 ffff842faf449c00 - [278063.225022] fe00: ffff842fb2a0fe30 ffff00000829d484 0000000000000020 ffff801fae2a9a00 - [278063.232925] fe20: 0000000025ccb150 ffff842fb2a0feb0 ffff842fb2a0fe70 ffff00000829ebbc - [278063.240827] fe40: ffff000009208000 ffff801fae2a9a00 ffff801fae2a9a00 0000000025ccb150 - [278063.248730] fe60: 0000000000000020 ffffffffffffffea 0000000000000000 ffff0000080838f0 - [278063.256632] fe80: 0000000000000000 0000801ff2faa000 ffffffffffffffff 0000ffffa7e06fe8 - [278063.264535] fea0: 0000000080000000 0000ffffa7e13128 00000000398a0000 0000000000040a01 - [278063.272437] fec0: 0000000000000003 0000000025ccb150 0000000000000020 0000000000000020 - [278063.280340] fee0: 0000000025ccb150 0000fffffe1023d8 00000000398a0000 0000000000000000 - [278063.288243] ff00: 000000000000003f 0000ffffa7e8e9a8 0101010101010101 0000000000000010 - [278063.296145] ff20: 000000000000000f ffffffffffffffff 0000000000000000 0000ffffa7eefcc0 - [278063.304048] ff40: 0000000000000000 000000000041c240 0000000000000a03 0000000000000020 - [278063.311950] ff60: 0000000025ccb150 0000000000000003 0000000000000000 0000000025ccb150 - [278063.319853] ff80: 000000000041c000 0000000000000000 000000000ee6b280 00000000000000fc - [278063.327755] ffa0: 0000000000000040 0000fffffe102410 000000000040622c 0000fffffe102410 - [278063.335658] ffc0: 0000ffffa7e06fe8 0000000080000000 0000000000000003 000000000000003f - [278063.343560] ffe0: 0000000000000000 0000000000000000 0000000000000000 0000000000000000 - [278063.351462] Call trace: - [278063.353982] Exception stack(0xffff842fb2a0fba0 to 0xffff842fb2a0fcd0) - [278063.360496] fba0: 0000000000000020 0001000000000000 000000000144d000 ffff0000084ce02c - [278063.368399] fbc0: 0000000060000145 ffff0000082b7600 ffff842fb3008540 ffff842faf449c00 - [278063.376301] fbe0: ffff842fb2a0fc10 ffff0000082c3f74 ffff842fb0131ca0 00000000ffffffff - [278063.384204] fc00: ffff842fb2a0fd30 0000000000000030 ffff842fb2a0fc30 ffff0000082a9b60 - [278063.392106] fc20: 0000000000000000 0000000000000038 ffff842fb2a0fc70 ffff0000082af37c - [278063.400008] fc40: 0000000000000000 ffff842fb2a0fd30 ffff842fb2a0fc70 0000000000040a01 - [278063.407911] fc60: 0000000025ccb150 ffff8000398a0000 0000000000000020 0000000000000020 - [278063.415813] fc80: 0000000000000000 0000000025ccb170 0000000025ccb150 0000000000000000 - [278063.423716] fca0: 000000000000003f 0000ffffa7e8e9a8 0101010101010101 0000000000000010 - [278063.431617] fcc0: 000000000000000f ffffffffffffffff - [278063.436568] [<ffff0000084ce02c>] __arch_copy_to_user+0xac/0x230 - [278063.442563] [<ffff00000829bf00>] __vfs_read+0x48/0x90 - [278063.447687] [<ffff00000829d484>] vfs_read+0x8c/0x148 - [278063.452725] [<ffff00000829ebbc>] SyS_read+0x6c/0xd8 - [278063.457677] [<ffff0000080838f0>] el0_svc_naked+0x24/0x28 - [278063.463063] Code: a8c12027 a88120c7 d503201f d503201f (a8c12027) - [278063.471015] systemd-journald[937]: Compressed data object 806 -> 576 using XZ - [278063.474566] ---[ end trace 69a9c073d1c44e35 ]--- - Segmentation fault (core dumped) + * Read from /dev/mem and scan DMI tables is dangerous, if /dev/mem is + not stable to read, it will cause lscpu crash. + + [Test Case] + + * `sudo lscpu` shall not cause Segmentation fault (core dumped) + + [Regression Potential] + + * Reading from DMI tables is sysfs is more stable then reading from + /dev/mem and if DMI tables is not available lscpu will use the old way + to search in /dev/mem. I believe its low regression risk
** Description changed: [Impact] - * Read from /dev/mem and scan DMI tables is dangerous, if /dev/mem is + * Read from /dev/mem and scan DMI tables is dangerous, if /dev/mem is not stable to read, it will cause lscpu crash. [Test Case] - * `sudo lscpu` shall not cause Segmentation fault (core dumped) + * `sudo lscpu` shall not cause Segmentation fault (core dumped) [Regression Potential] - * Reading from DMI tables is sysfs is more stable then reading from + * Reading from DMI tables is sysfs is more stable then reading from /dev/mem and if DMI tables is not available lscpu will use the old way to search in /dev/mem. I believe its low regression risk -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1674258 Title: lscpu crashs when /dev/mem is not stable to read To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/util-linux/+bug/1674258/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
